Referee 2's review:

**********
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication, the manuscript has been revised in light of these comments and to address other editorial requirements.
**********

The authors describe a new type of assay for BoNT/A detection that combines an immunocapture and the measurement of the metalloprotease activity of the toxin. They claim the assay reaches attomolar sensitivities in complex matrices such as serum and milk, thus being considerably more sensitive and faster than the mouse bioassay. The performances of this method should make it attractive for diagnostic, biodefense and pharmacological applications.

The description of the method is detailed and the authors place well their claims in the context of previous literature by providing an adequate review of the performances of major existing botulinum toxin assays.

However, and this is the main critical point, the only serum data presented are in 10% fetal bovine serum (FBS). This is surprising, and one could wonder whether it means the assay does not perform well in human serum ? This point is important because it is in human serum, not so much in milk and not in FBS, that one needs very high sensitivities for a BoNT detection assay. Therefore, to be truly convincing in supporting their claims, the authors should provide additional data showing that sensitivity is retained in human serum samples (the use of human serum obtained from pooled donors, available from Sigma Chemical or other suppliers, would be recommended to account for interfering antibodies found in many human sera).